Skip to main content

​
Gemini Provider

The Gemini provider gives you access to Google’s Gemini models, featuring the efficient Flash series and powerful Pro models with extensive context windows.

​
Configuration

​
API Key Setup

Set your Google API key in your .env file: 
GEMINI_API_KEY=your_api_key_here
The API key configuration is defined in config/llm-magic.php: 
'apis' => [
    'gemini' => [
        'token' => env('GEMINI_API_KEY'),
    ],
    'google' => [
        'token' => env('GEMINI_API_KEY'),
    ],
]

​
Available Models

LLM Magic supports the following Gemini models:

​
Gemini 1.5 Series

  • gemini-1.5-flash - Fast and efficient model
  • gemini-1.5-flash-8b - Compact 8B parameter model
  • gemini-1.5-pro - High-performance model with 2M token context

​
Gemini 2.0 Series

  • gemini-2.0-flash - Latest Flash model optimized for agents
  • gemini-2.0-flash-lite - Lightweight Flash variant

​
Gemini 2.5 Series

  • gemini-2.5-flash-lite - Compact Flash model
  • gemini-2.5-flash - Latest Flash model
  • gemini-2.5-pro - Latest Pro model

​
Experimental Models

  • gemini-2.5-flash-preview-04-17 - Flash preview build
  • gemini-2.5-pro-preview-03-25 - Pro preview build
  • gemini-2.5-pro-exp-03-25 - Pro experimental build

​
Model Constants

Use these constants for type safety: 
use Mateffy\Magic\Models\Gemini;

Gemini::GEMINI_1_5_FLASH             // gemini-1.5-flash
Gemini::GEMINI_1_5_FLASH_8B          // gemini-1.5-flash-8b
Gemini::GEMINI_1_5_PRO               // gemini-1.5-pro
Gemini::GEMINI_2_0_FLASH             // gemini-2.0-flash
Gemini::GEMINI_2_0_FLASH_LITE        // gemini-2.0-flash-lite
Gemini::GEMINI_2_5_FLASH_LITE        // gemini-2.5-flash-lite
Gemini::GEMINI_2_5_FLASH             // gemini-2.5-flash
Gemini::GEMINI_2_5_PRO               // gemini-2.5-pro
Gemini::GEMINI_2_5_FLASH_PREVIEW     // gemini-2.5-flash-preview-04-17
Gemini::GEMINI_2_5_PRO_PREVIEW       // gemini-2.5-pro-preview-03-25
Gemini::GEMINI_2_5_PRO_EXPERIMENTAL  // gemini-2.5-pro-exp-03-25

​
Usage

​
Using the Constructor

Create a Gemini model instance: 
use Mateffy\Magic\Models\Gemini;
use Mateffy\Magic\Models\Options\ChatGptOptions;

$model = new Gemini(
    model: Gemini::GEMINI_2_0_FLASH,
    options: new ChatGptOptions
);

​
Using Static Factory Methods

LLM Magic provides convenient static methods: 
use Mateffy\Magic\Models\Gemini;

// Flash models (defaults to latest version)
$model = Gemini::flash();        // Uses flash_2
$model = Gemini::flash_1_5();
$model = Gemini::flash_2();
$model = Gemini::flash_2_lite();
$model = Gemini::flash_2_5_flash();

// Pro models
$model = Gemini::flash_2_5_pro();

// Preview/Experimental
$model = Gemini::flash_2_5_flash_preview();
$model = Gemini::flash_2_5_pro_preview();
$model = Gemini::flash_2_5_pro_experimental();

​
Getting Available Models

Retrieve a list of all available Gemini models: 
use Mateffy\Magic\Models\Gemini;

$models = Gemini::models();
// Returns a Collection with prefixed model names like 'google/gemini-2.0-flash'

$models = Gemini::models(prefix: null, prefixLabels: null);
// Returns models without prefix

​
Model Costs

Gemini models include built-in cost tracking. The following pricing is configured (per 1M tokens):
ModelInput CostOutput Cost
Gemini 1.5 Flash$0.15$0.60
Gemini 1.5 Flash 8B$0.075$0.30
Gemini 1.5 Pro$2.50$10.00
Gemini 2.0 Flash$0.10$0.40
Gemini 2.0 Flash Lite$0.075$0.30
Gemini 2.5 Flash Preview$0.15$0.60
Gemini 2.5 Pro Preview$1.25$5.00
Gemini 2.5 Pro Experimental$1.25$5.00
Access cost information programmatically: 
$model = Gemini::flash_2();
$cost = $model->getModelCost();

if ($cost) {
    $inputCostPerMillion = $cost->inputCentsPer1K * 1000;
    $outputCostPerMillion = $cost->outputCentsPer1K * 1000;
}

​
API Configuration

Gemini uses the OpenAI-compatible API with a custom base URI: 
protected function getOpenAiBaseUri(): ?string
{
    return 'generativelanguage.googleapis.com/v1beta/openai';
}

​
Organization Info

The Gemini provider includes organization metadata:
  • ID: google
  • Name: Google
  • Website: https://google.com
  • Privacy: Data may be used for model training and abuse prevention

​
Features

​
Multimodal Support

Gemini models support text, image, video, and audio inputs:
  • Text: All models
  • Images: All models
  • Video: Flash and Pro models
  • Audio: 2.0 Flash and newer

​
Large Context Windows

  • Gemini 1.5 Pro: Up to 2 million tokens
  • Gemini 2.0 Flash: 1 million tokens
  • Gemini 2.5 Flash: 1 million tokens

​
Context Caching

Gemini 2.0 Flash and later support context caching for reduced costs on repeated queries.

​
Advanced Options

You can pass additional options using ChatGptOptions: 
use Mateffy\Magic\Models\Gemini;
use Mateffy\Magic\Models\Options\ChatGptOptions;

$options = new ChatGptOptions(
    // Configure temperature, max tokens, etc.
);

$model = new Gemini(Gemini::GEMINI_2_0_FLASH, $options);

​
See Also

Build docs developers (and LLMs) love

Get started for freeTalk to us